The PCNA interaction protein box sequence in Rad54 is an integral part of its ATPase domain and is required for efficient DNA repair and recombination

Abstract

Rad54 is an ATP-driven translocase involved in the genome maintenance pathway of homologous recombination (HR). Although its activity has been implicated in several steps of HR, its exact role(s) at each step are still not fully understood. We have identified a new interaction between Rad54 and the replicative DNA clamp, proliferating cell nuclear antigen (PCNA). This interaction was only mildly weakened by the mutation of two key hydrophobic residues in the highly-conserved PCNA interaction motif (PIP-box) of Rad54 (Rad54-AA). Intriguingly, the rad54-AA mutant cells displayed sensitivity to DNA damage and showed HR defects similar to the null mutant, despite retaining its ability to interact with HR proteins and to be recruited to HR foci in vivo. We therefore surmised that the PCNA interaction might be impaired in vivo and was unable to promote repair synthesis during HR. Indeed, the Rad54-AA mutant was defective in primer extension at the MAT locus as well as in vitro, but additional biochemical analysis revealed that this mutant also had diminished ATPase activity and an inability to promote D-loop formation. Further mutational analysis of the putative PIP-box uncovered that other phenotypically relevant mutants in this domain also resulted in a loss of ATPase activity. Therefore, we have found that although Rad54 interacts with PCNA, the PIP-box motif likely plays only a minor role in stabilizing the PCNA interaction, and rather, this conserved domain is probably an extension of the ATPase domain III.

Figure 1. The Rad54 family contains a conserved PCNA interaction motif, and yeast Rad54 directly interacts with PCNA.

A. Alignment of the Rad54 family of proteins. The conservation of the amino acid sequence around motif III in Rad54 family members from budding yeast to human is depicted. Amino acids shaded in black are identical between the sequences, while grayed amino acids retain highly similar side chains. The location of the PCNA interaction protein box (PIP-box) is shown below the alignment. The two C-terminal highly-conserved aromatic residues were changed to alanines by site-directed mutagenesis to create the Rad54-AA (Y494A, F495A) mutant protein. B. Section of the Rad54 structure containing the PIP box and ATPase domains. The tertiary structure of the protein Rad54 from D. rerio (Zebrafish (PDB ID 1Z3I)) is represented in the ribbon diagram. Motif III (317-ISGTPIQN-324, corresponding to amino acids 481–489 in S. cerevisiae Rad54) is shown in blue, the PIP-box motif (323-QNDLLEYF-330, corresponding to amino acids 488–495 in S. c. Rad54) is shown in red, while overlapping residues assigned to both motifs (323-QN-324) are shown in magenta. The detailed view (inset) shows these two motifs and mutated residues Y329 (494) and F330 (495) in stick representation. C. Rad54 interacts with PCNA in vitro. Rad54 protein was mixed with either control beads (affi-BSA, lanes 1–3) or with PCNA immobilized on affi-beads (lanes 4–6). After incubation for 30 min at 4°C with occasional mixing, the reaction mixtures were centrifuged and the supernatant was separated from the beads. Next, the beads were washed and the proteins were eluted by SDS-PAGE loading dye. Supernatant (S), wash (W) and eluate (E) fractions were separated on a 12% SDS-PAGE gel, followed by blotting and detection with α-Rad54 antibodies. D. Oligopeptides derived from the PIP-box outcompete PCNA for interaction with Rad54. In the pull-down experiments shown, Rad54 was mixed with immobilized PCNA in the presence of an oligopeptide derived from the Rad54 PIP box (pFF; lanes 2, 5) or its mutated version (pAA; lanes 3, 6). Lanes 1, 4 represent control experiment where no peptide was added. The reactions were performed as described in C. Supernatant (S), and eluate (E) fractions were separated on a 12% SDS-PAGE gel, followed by Coomassie staining. E. PCNA interaction with Rad54 and Rad54-AA. In these pull-down experiments, Rad54 (lanes 6 and 8) or Rad54-AA (lanes 7 and 9) were mixed with PCNA immobilized on affi-beads, or with BSA affi-beads as a control (lanes 4, 5). The reactions were performed as described in C. Input (I) and eluate (E) fractions were separated on a 12% SDS-PAGE gel, followed by Coomassie staining.

Figure 2. The Rad54 PIP-box mutant renders cells defective in homologous recombination and genome stability functions.

A. Cells with the rad54-AA mutation are as sensitive to DNA damaging agents as the null mutant. Shown are 10-fold dilutions of cells spotted onto plates treated with methylmethane sulfonate (MMS), hydroxyurea (HU) or ultraviolet (UV) light, at the indicated doses. The spot assay was performed as described in the methods. B. The rad54 PIP-box mutant is defective in mitotic recombination and genome stability functions. Results from wild type, rad54Δ and rad54-AA mutant strains in recombination assays are shown, and the fold change from the wild type is shown in parentheses. Gene conversion and deletion (single-strand annealing) event rates were determined by fluctuation tests with the leu2-EcoRI::URA3-leu2-BstEII reporter as described in the methods. The mean of the rates from three independent experiments are shown with standard deviations. Spontaneous mutagenesis rates in the rad54-AA mutant were determined two or three times by fluctuation tests, as described. Significance was determined using a t-test (p<0.05). For spore viability, diploids homozygous for RAD54, rad54Δ, or rad54-AA were sporulated and dissected, and the surviving spores quantified. At least 100 spores were analyzed for each strain.

Figure 3. Rad54-AA is defective in strand invasion and primer extension activities.

A. Schematic of the MAT chromosomal locus used for the examination of DNA repair synthesis. Arrows depict the direction of primers used for detection of primer extension intermediates by PCR. B. The primer extension step of recombination is compromised in the rad54 PIP-box mutant. The top panel shows the formation pA-pB product, which results from minimal DNA synthesis from the invading strand. Samples were taken at 1, 2 or 5 h after HO endonuclease cutting. The bottom panel shows pC-pF control product. C. Rad54-AA is defective in DNA repair synthesis in vitro. Rad51 and DNA substrates were pre-formed into nucleoprotein filaments as described, then either Rad54 wild type (wt, lanes 1–7) or Rad54-AA (lanes 8–14) was incorporated and D-loop formation was initiated. DNA synthesis reactions were then performed using Polymerase δ (15 nM), and increasing concentrations (2.5, 5, 10, 20 nM) of the PCNA clamp, with or without the PCNA clamp loader, RFC (10 nM), in the presence of RPA (666 nM). The reactions were monitored using labeled α-[³²P]-dATP, and percentage of each reaction product shown below.

Figure 4. The rad54-AA mutant protein is deficient in most of its biochemical activities.

A. Rad54-AA has lower ATPase activity compared to wild type Rad54. Rad54-AA and Rad54 wt (75 nM, each), respectively, were mixed with dsDNA and α-[³²P]-labeled ATP. At indicated times, samples were withdrawn and analyzed by thin-layer chromatography. Error bars represent standard error produced by 3 experiments. B. Rad54-AA does not branch migrate mobile Holliday junctions. DNA substrate was incubated with increasing concentrations (2.5, 5, 10, 20 nM) of Rad54 wt (lanes 2–5) or Rad54-AA (lanes 6–9), respectively, in the presence of ATP. Lane 1 shows the no protein control reaction.

Figure 5. Rad54-AA cannot bind dsDNA as efficiently as wild type Rad54.

A. Rad54-AA performs less well than the wild type in a DNA binding assay in the absence of ATP. Purified S. cerevisiae Rad54 and Rad54-AA (31.25, 62.5, 125, 250, 500 or 1000 nM) were incubated for 10 min with linearized pBluescript plasmid to assess DNA binding. Prior to gel electrophoresis, the proteins were cross-linked to DNA with 0.1% glutaraldehyde. After the addition of gel loading buffer, the reaction mixtures were resolved in a 0.8% agarose gel in TAE buffer and stained with Midori Green DNA stain. B. Rad54-AA performs less well than the wild type in a DNA binding assay in the presence of ATP. DNA binding assay as performed exactly as in A, except for the addition of 2.5 mM ATP, and an ATP-regenerating system to the reaction. C. Quantification of the DNA binding reactions shown in A and B. Error bars represent the standard error from three independent trials.